This demonstration shows how malicious content in a CV can manipulate an AI hiring agent to exfiltrate data or influence evaluation scores.
Your company uses an AI hiring agent to evaluate candidate CVs. The agent has access to internal salary data and makes hiring recommendations. Candidates upload their CVs directly to the system.
Select the CV you want the hiring agent to evaluate. The benign CV represents a trustworthy applicant, while the malicious CV embeds hidden instructions to exfiltrate data.